BACKGROUND: Veterans with diabetes must control cardiovascular risk factors in order to prevent disabling and life-threatening complications. However, despite system wide advances in diabetes quality of care, over 30% of VHA patients with diabetes continue to have uncontrolled blood pressure, hyperglycemia, or hyperlipidemia. The nationwide VA PACT (Patient-Aligned Care Teams) initiative seeks to provide VA primary care patients with comprehensive, team-based support for following diabetes care regimens. PACT's success, however, hinges on its ability to effectively engage patients in care. One relatively untapped resource for supporting engagement in PACT is patients' family and friends. Three out of four adults with diabetes reach out to an unpaid family member or friend (a `health supporter') for ongoing help with diabetes management. These supporters help patients with medication adherence, tracking home glucose measurements, maintaining a healthy eating plan, and often accompany patients to their medical visits. However, while PACT emphasizes the importance of family members as part of the care team, PACT does not have formal mechanisms to involve health supporters in PACT care. Health supporters report that, in order to be more effective, they need more information on patient's medical care plans, clear channels for communicating with PACT team members, and information on navigating PACT resources. OBJECTIVES: The overall objective of this randomized trial is to test a strategy to strengthen the capacity of supporters to help patients with high-risk diabetes engage in PACT care and successfully enact care plans. The central hypothesis is that providing health care engagement tools to both health supporters and patients will increase patient activation and improve management of diabetes complication risks. METHODS: This will be a randomized controlled trial evaluating an intervention (Caring Others Increasing EngageMent in PACT, or CO-IMPACT) designed to structure and facilitate health supporter involvement in PACT so that patients can become more actively engaged in PACT care, and improve their diabetes treatment processes and outcomes. 220 patients with diabetes receiving PACT primary care who 1) are at high risk for diabetes complications due to hyperglycemia OR high blood pressure and 2) have a health supporter involved in their diabetes care will be recruited along with their health supporter. Patient-supporter dyads will be randomized to the CO-IMPACT intervention or usual PACT care for high-risk diabetes, for 12 months. The CO- IMPACT protocol provides patient-supporter dyads: one coaching session on action planning, communicating with providers, navigation skills and support skills; preparation by phone before patients' primary care visits; after-visit summaries by mail; and biweekly automated phone calls to prompt dyad action on new patient health concerns. CO-IMPACT builds on medical record-integrated patient activation tools in the PACT toolkit and is designed to be implementable within existing PACT nurse encounters. Primary outcomes for this study include a validated measure of patient activation (Patient Activation Measure-13) and a cardiac event 5-year risk score designed for patients with diabetes (UKPDS Risk Engine). Secondary outcomes include patients' self- efficacy for diabetes self-care; diabetes self-management behaviors including medication adherence; diabetes distress; and glycemic and blood pressure control. Measures among supporters will include supporter activation, use of effective support techniques, distress about patient's diabetes care, and caregiver burden. We will also measure patient-supporter and patient-provider relationship quality, patient safety (e.g. hypoglycemia), and utilization. We will measure potential moderators of intervention effect, such as patient health literacy level, and facilitators and barriers to wider implementation among participants and staff.